模型bug修复

spark/featureEngSk.py

@@ -967,7 +967,7 @@ if __name__ == '__main__':
     itemStatisticSql = getItemStatisticSql(itemStatisticStartDay, endDay)
 
     itemStatisticDF = spark.sql(itemStatisticSql)
-    itemStatisticDF.show(100, False)
+    # itemStatisticDF.show(100, False)
 
     partitionDatas = generatePartitionDates(trainDays + itemStatisticStartDays)
     partitionDatasBC = spark.sparkContext.broadcast(partitionDatas)
@@ -983,7 +983,18 @@ if __name__ == '__main__':
         return res
 
     itemStatisticDF = itemStatisticDF.rdd.flatMap(splitPatitionDatasFlatMapFunc).toDF(["card_id", "partition_date", "label", "label_count"])
-    itemStatisticDF.orderBy(['card_id', 'label', 'partition_date']).show(1000, False)
+    itemStatisticDF.orderBy(['card_id', 'label', 'partition_date'])
+    itemStatisticDF.createOrReplaceTempView("itemStatisticDF")
+    itemStatisticSql = """
+    SELECT 
+        card_id, 
+        partition_date, 
+        label,
+        SUM(label_count) OVER(PARTITION BY card_id ORDER BY partition_date ROWS BETWEEN ${itemStatisticStartDays} PRECEDING AND 1 PRECEDING) label_count_sum
+    FROM
+        itemStatisticDF
+    """.format(itemStatisticStartDays = itemStatisticStartDays)
+    spark.sql(itemStatisticSql).show(100, False)
     sys.exit(1)
     clickDF = spark.sql(clickSql)
     clickDF.createOrReplaceTempView("clickDF")